Field of the Invention
The subject matter described herein relates generally to diet management and nutrition analysis. More specifically, the present disclosure is related to computer vision techniques of food recognition based on an image of the food and the food container.
Description of Related Art
The lack of accurate weight information in nutrition analysis and inaccuracy of food recognition can lead to inefficient diet management. Currently, there exist some handheld applications and computer vision systems that capture food images for nutritional value analysis. Due to many varying factors, such as lack of a predefined container, lighting, focal length, which exists in the conventional food image recognition methods, the accuracy of the nutrition analysis is very low. One of the reasons these existing image recognition technologies performs poorly when assessing accurate nutrition data is largely due to their inability to capture volume, density, and/or weight of the food from just a single image of the food. Volume, density, and weight are difficult to gauge without a predefined container included in the image. There are currently no solutions that provide accurate identification and weight information of a food item due to the absence of a fixed and predefined container which is needed for reference. Moreover, there are no solutions currently available to provide accurate identification and weight information from the combination of a fixed and predefined container that has a load sensor to communicate the exact weight of the food item.
Many research studies have recently shown clear correlations between poor dieting and health conditions. Poor diets, such as overeating, under eating, and/or consuming poor quality foods have been linked to certain medical conditions like diabetes, cancer, heart disease, hypertension, chronic kidney disease, obesity, and the like. For consumers and patients trying to control the quality and habit of consuming food, it is often difficult to identify what kinds of food they are consuming and to improve their eating habits and health conditions. It is cumbersome for the consumers to meet their prescribed nutritional requirements or dietary regimen necessary each time they prepare a meal.
Therefore, there is a need for a system and method that accurately recognizes the type of food being consumed and the weight of the food, in order to be aware of the nutrition and calories being consumed at each meal. In addition, there also is a need for an accurate computer vision technique system and method that can identify the food being consumed at each meal.